Successful Automated Normalization of Cancer Outcomes for Half a Million Patients Across Four Disparate Health Systems.

e18763Background: The lack of interoperability among different electronic medical record (EMR) systems remains a challenge to fulfilling the promises of precision oncology. Typically, cancer outcomes data is stored as unstructured information in a multitude of inconsistent fields across EMRs. This prevents the integration of outcomes into the treatment decision-making process. To overcome this challenge, we leveraged the Syapse ontology, a data model that unifies biomedical data. We report our successful efforts to unify outcomes data across a large network of health systems. Methods: Health system participants included Providence St. Joseph Health, Aurora Healthcare, Henry Ford Health System, and University of Miami Sylvester Comprehensive Cancer Care. Data integration efforts were initiated with each health system between 1/2015 and 1/2018. Syapse aggregated clinical data sources and automated outcomes calculation. Source datasets within the health system included the EMR, enterprise data warehouse, tum...